Development of new anti-cancer drugs is becoming prohibitively expensive, even more so as therapeutics become more targeted against sub-populations of patients. It is hypothesized that non-invasive imaging biomarkers can be used to monitor and predict responses to anti-cancer drugs and, hence, improve efficiency of drug development. Many potential imaging biomarkers exist and a significant barrier to their use is simply making the right choice for each drug-cancer combination. The current project seeks to reduce the dimensionality of this problem by investigating multiple magnetic resonance imaging (MRI) endpoints to targeted therapies in animal models. This work is predicated on the explicit assumption that these results can empirically and rationally determine appropriate imaging endpoints in a clinical trial setting. Prior work by our group and others has identified a number of useful MR imaging endpoints for drug therapy, including diffusion (DW) MRI, dynamic contrast enhanced (DCE) MRI, 1H magnetic resonance spectroscopy (MRS) and blood oxygen level dependent (BOLD) MRI. In animals, these can all be obtained in a single imaging session and will be used to monitor the effect of PX-866, a novel inhibitor of phosphatidylinositol-3-kinase (PI3K) and sunitinib, a small molecule inhibitor of receptor tyrosine kinases, VEGF-R and PDGF-R. These will be used against a panel of breast cancer tumor xenografts with differential sensitivities to these drugs. Primary goals of this work are to: (1) determine which of these MR imaging biomarkers are most useful for each drug; and (2) whether imaging biomarker responses can differentiate between these drug classes. Aim 1 will investigate this by assessing tumor and molecular responses to PX-866 and sunitinib. Aim 2 will assess the imaging responses to these drugs to determine which imaging endpoints are most appropriate as response biomarkers and Aim 3 will retrospectively analyze the pre-therapy images to determine if any biomarker (or combination) would have been useful in predicting response. This work will advance our knowledge by: (1) determining which imaging biomarkers are most appropriate for these cancer models and these drugs; (2) identifying imaging phenotypes that may be predictors of response; and (3) developing novel and more sensitive methods of image analysis. PUBLIC HEALTH RELEVANCE: Diagnostic radiology is undergoing a revolution as it transforms from a discipline focused on anatomy to one that can non-invasively interrogate tissue function and molecular phenotypes. There is strong interest to develop imaging biomarkers in clinical trial settings to monitor and predict responses to anti-cancer drugs. The current project seeks to further this development by testing the hypothesis that magnetic resonance imaging (MRI) endpoints can monitor and/or predict the response of breast cancers to targeted therapies. The current work will be performed in animal models with the explicit assumption that these results can empirically and rationally determine which imaging endpoints will be optimal in a clinical setting. This work will use a panel of .breast cancer cell lines whose molecular phenotypes represent the heterogeneity of this disease.